Gender-Specific Costs of Reproduction on Vegetative Growth And

Annals of Botany 106: 989–998, 2010 doi:10.1093/aob/mcq197, available online at www.aob.oxfordjournals.org

Gender-speciﬁc costs of reproduction on vegetative growth and physiological performance in the dioecious shrub Corema album

Leonor A´ lvarez-Cansino*, Marı´a Zunzunegui, Mari Cruz Dı´az Barradas and Mari Paz Esquivias Department of Plant Biology and Ecology, University of Seville, Apt. 1095, 41080, Seville, Spain * For correspondence. E-mail [email protected]

Received: 12 July 2010 Returned for revision: 29 July 2010 Accepted: 31 August 2010 Published electronically: 29 September 2010 Downloaded from https://academic.oup.com/aob/article-abstract/106/6/989/140671 by guest on 05 May 2020

† Background and Aims Reproductive costs imply trade-offs in resource distribution at the physiological level, expressed as changes in future growth and/or reproduction. In dioecious species, females generally endure higher reproductive effort, although this is not necessarily expressed through higher somatic costs, as compensatory mechanisms may foster resource uptake during reproduction. † Methods To assess effects of reproductive allocation on vegetative growth and physiological response in terms of costs and compensation mechanisms, a manipulative experiment of inflorescence bud removal was carried out in the sexually dimorphic species Corema album. Over two consecutive growing seasons, vegetative growth patterns, water status and photochemical efficiency were measured to evaluate gender-related differences. † Key Results Suppression of reproductive allocation resulted in a direct reduction in somatic costs of reproduction, expressed through changes in growth variables and plant physiological status. Inflorescence bud removal was related to an increase in shoot elongation and water potential in male and female plants. The response to inflorescence bud removal showed gender-related differences that were related to the moment of maximum reproductive effort in each sexual form: flowering in males and fruiting in females. Delayed costs of reproduction were found in both water status and growth variables, showing gender-related differences in resource storage and use. † Conclusions Results are consistent with the existence of a trade-off between reproductive and vegetative biomass, indicating that reproduction and growth depend on the same resource pool. Gender-related morphological and physiological differences arise as a response to different reproductive resource requirements. Delayed somatic costs provide evidence of gender-related differences in resource allocation and storage. Adaptive differences between genders in C. album may arise through the development of mechanisms which compensate for the cost of reproduction.

Key words: Corema album, costs of reproduction, ﬂower removal, photochemical efﬁciency, sexual dimorphism, trade-off, water potential.

INTRODUCTION traits, such as stem growth and flower and fruit production (Matsuyama and Sakimoto, 2008). The reproductive costs hypothesis states that plant resource Nevertheless, reproduction may not always incur a detect- allocation or development is limited such that an increase in able cost, because compensatory mechanisms may exist to ongoing plant reproduction incurs a reduction in fitness, quan- foster resource uptake during reproduction (Tuomi et al., tified as reduced future fecundity or survival (Levins, 1968; 1983), rendering these intersexual differences in reproductive Karlsson and Meńdez, 2005). The main assumption in the trade-offs non-detectable. Several mechanisms which compen- hypothesis of reproductive costs is that a trade-off occurs in sate for higher reproductive allocation in females have been the distribution of resources at the physiological level, either described in dioecious species; these include (a) different towards vegetative growth or towards reproduction. Thus, it timing of vegetative and reproductive allocation (Delph, could be expected that plants express physiological costs of 1990; Obeso, 2002; Milla et al., 2006), (b) microhabitat parti- reproduction through a reduction in growth or in future repro- tioning among the sexes (Bierzychudek and Eckhart, 1988; duction (Fox, 1995). Dawson and Geber, 1999) and (c) increased photosynthetic Dioecious plants, in which reproductive functions are separ- activity during reproduction (Dawson and Ehleringer, 1993; ated in two distinct sexual forms, are especially suitable for Obeso, 2002; Nicotra et al., 2003). studying interactions between growth and reproduction, as In this context and assuming equal conditions, female indi- the effects of flower and fruit production can be analysed sep- viduals should have greater capacity than males for resource arately (Hoffmann and Alliende, 1984). In dioecious woody acquisition through uptake, assimilation and/or reabsorption species, the timing and extent of the reproductive cost are throughout the reproductive period, but particularly during reported to differ between sexual forms, with a general fruiting, when greater resource allocation occurs (Ashman, pattern of females enduring a higher net total cost than 1994; Delph, 1999). males (Leigh et al., 2006). Intersexual differences in reproduc- Sequencing of vegetative growth and fruiting, or non- tive allocation can result in relative differences in life-history overlapping phenophases, has been described to be a

# The Author 2010. Published by Oxford University Press on behalf of the Annals of Botany Company. All rights reserved. For Permissions, please email: [email protected] 990 A´ lvarez-Cansino et al. — Gender-specific costs of reproduction in a dioecious shrub compensating mechanism, reducing competition for resources relations and photosynthetic assimilation would be affected by between reproductive and vegetative organs (Delph, 1990; the different reproductive demands between genders. Biomass Matsuyama and Sakimoto, 2008). Differences in the timing allocation pattern was also measured, as this variable could be of vegetative growth and reproduction have been described related to an increase in photosynthetic assimilation related to to vary between sexual forms in an array of woody dioecious reproductive costs of both current and ensuing reproductive species (A˚ gren, 1988; Popp and Reinartz, 1988; Matsuyama seasons (Case and Barrett, 2004; Milla et al.,2006). and Sakimoto, 2008). In Pistacea lentiscus, Milla et al. (2006) described the sequencing of flowering and growing periods to explain the lack of trade-offs between reproduction MATERIALS AND METHODS and vegetative growth, diminishing reproductive costs. In the Study species dioecious shrub Corema album, Zunzunegui et al. (2006) found evidence of gender differences in the timing of Corema album (Ericaceae) is an evergreen dioecious shrub Downloaded from https://academic.oup.com/aob/article-abstract/106/6/989/140671 by guest on 05 May 2020 maximum vegetative growth and reproductive investment which rarely exceeds 1 m in height. The species is also con- within a growing season. Early growth in females has been sidered as subdioecious because male plants occasionally explained by several authors as a compensation for their present inflorescences with staminate flowers only and inflores- greater resource needs for fruit production, allowing a cences with both pistillate and staminate flowers, although her- greater allocation of resources to storage organs and/or the pro- maphrodite individuals are scarce (1–4 %) and only present at duction of a greater leaf area before the subsequent reproduc- the southern most populations of the species’ geographical tive event (Popp and Reinartz, 1988; Delph, 1990; Obeso, range (Zunzunegui et al., 2006). Leaves are ericoid, alternated 2002; Matsuyama and Sakimoto, 2008). or verticillated and persist for only two growing seasons. Manipulative experiments have been used to assess balances Pollination is anemophilous and fruits are spherical white or in resource allocation and reproductive costs, as they constitute pink–white berries (5–8 mm in diameter). The growing a fixed environmental effect independent of the genetic origin, period takes place from February to July, reaching its and therefore allows isolation of the indirect demographic maximum between April and June, while flowering occurs costs of adaptations that compensate for reproductive costs from February to April, with fruits ripening from June to (Hartemink et al., 2004; Andersson, 2005; Gehring and July. Corema album is endemic to the Atlantic coast of the Delph, 2006; Narbona and Dirzo, 2010; among others). Iberian Peninsula, from north-west Spain to the proximities However, experiments employing direct manipulation of of the Straits of Gibraltar in the south, growing on sandy reproduction may be limited in terms of describing trade-offs, soils over coastal dunes and cliffs (Valde´s et al., 1987). if storage organs such as stems or roots compete with repro- Previous studies have reported that reproductive effort is ductive structures as sinks altering the sink : resource ratio 3-fold higher in female than in male plants in terms of (allocation to reproductive vs. photosynthetic tissues). This biomass at different populations along the geographical has been described as being of particular importance in seaso- range of C. album: Guitiań et al. (1997) in the north of the nal habitats (Ehrleń and van Groenendael, 2001). Thus, studies species range and Zunzunegui et al. (2006) in the south at spanning at least two reproductive events are considered to be the same study site selected for the present study. more accurate in detecting the effects of manipulations of current reproductive allocation on future growth and/or repro- Study site duction (Primack and Hall, 1990; Ehrleń and van Groenendael, 2001; Garcıá and Ehrleń, 2002; Horibata et al., 2007). In The study was carried out in the coastal dunes of El Asperillo addition, very few studies have analysed the effects of repro- (34 80’N 6 836’W), in the Donãna Natural Park (south-west ductive allocation on growth and physiological response con- Spain). The climate is Mediterranean type with oceanic influ- comitantly (but see Dumka et al., 2003; Horibata et al., 2007). ence and mild temperatures. Mean annual temperature is 16.8 In the present study, a direct manipulation experiment invol- 8C. Average annual rainfall is 550 mm, concentrated between ving inflorescence bud removal was performed with the aim of October and March. Vegetation consists of native dune scrub analysing reproductive costs on vegetative growth and physio- of Halimium halimifolium, H. commutatum, Cistus libanotis, logical response of male and female Corema album individ- C. salvifolius, Rosmarinus officinalis, Lavandula stoechas, uals, determining the effects of current reproductive Cytisus grandiflorus, Stauracanthus genistoides, Corema allocation over two consecutive years. In particular, the fol- album, Juniperus oxycedrus and J. phoenicea,togetherwith lowing questions were addressed. (a) Are sexual forms differ- Pinus pinea. Corema album is the dominant scrub in some ent in terms of trade-offs between reproduction and vegetative areas (Dıáz Barradas and Munõz Reinoso, 1992). growth? (b) Is there a physiological mechanism compensating The study was conducted from February to July over two for reproductive costs? (c) Does reproduction cause indirect consecutive growing periods, 2004 and 2005. The second (delayed) somatic or physiological costs? A greater effect of study period was especially dry, with an annual rainfall 60 % inflorescence bud removal on vegetative growth and physi- below the historic yearly average for the area (179 mm). ology on female individuals was expected, as this is the sexual form with the greatest reproductive allocation in Experimental design C. album (Guitiań et al., 1997; Zunzunegui et al., 2006). To test the existence of physiological compensatory mechan- Ten even-sized individuals of each gender were randomly isms, leaf water potential and photochemical efficiency of chlor- chosen and marked within a 50 × 50 m area, where male ophyll a were determined, as it was anticipated that water and female canopies show a random spatial distribution and A´ lvarez-Cansino et al. — Gender-specific costs of reproduction in a dioecious shrub 991 the sex ratio is equilibrated (A´ lvarez-Cansino, 2009). At the dates divided by the number of days elapsed between them: beginning of 2004, all inflorescence buds were removed, on =[ − ] [ − ]() initial appearance and before flowers could bloom (in SErate SEt SE(t−1) / dayst days(t−1) 1 January for male individuals and in February for females), with the aim of suppressing reproduction. Ten other similar Accumulated shoot elongation at the end of the growing period individuals of each gender were marked for control purposes (SEa, mm) was measured as the final SE value accumulated per in the same area. All selected individuals were reproductive shoot at the end of each study year (2004 and 2005). in the two study years and no hermaphrodite flowers were The average number of shoots produced per measured detected on any of them. branch was calculated for an estimate of ramification per Gender determination was performed by searching for individual. remains of anthers and fruits from the previous reproductive Mean shoot elongation per shoot (MSE, mm) through the period. Residual late inflorescences were systematically whole growing season was estimated as the ratio between Downloaded from https://academic.oup.com/aob/article-abstract/106/6/989/140671 by guest on 05 May 2020 removed over a 2-week period after the start of the study. SE and the number of ramifications per terminal shoot Measurements of major (A) and minor (a) axes (A orthog- (eqn 2). onal with respect to a) of canopy projection were carried out on each selected individual. Canopy area was calculated MSE = SE/#ramifications (2) approaching canopy projection to an ellipse. The range of canopy size overlapped between sexes (range of mean values of A and a:0.5–1.025 m for males and 0.49–1.03 m for Biomass production females), with no significant differences between genders (ANOVA, F1 ¼ 0.663, P ¼ 0.426). To determine biomass production per shoot, at the end of To estimate the number of flowers produced per plant, all the experiment all the new shoots produced in the 2005 inflorescence buds removed at the beginning of the experiment growing season from all branches marked in 2004 were cut (2004) were counted and the number of flowers per square at the level of the last bud-scale scar, which was produced metre of canopy was calculated with the area projection of during the 2005 flowering event (n ¼ 400, 10 branches per each individual. 10 individuals per 2 gender per 2 treatment). After drying in To estimate the number of fruits produced per plant, and a forced-air oven at 80 8C for 24 h, plant material was separ- account for inter-annual variation in fruit production, two ated into leaves, stems and fruits (in 2005, buds were not 0.25 m2 quadrats were randomly placed in the canopy of removed, hence, fruits for both treatments were available) each control plant when fruits had ripened (July) in the 2004 and weighed for the estimation of relative biomass devoted and 2005 reproductive events. All fruits within the quadrats to fruit production as: were counted and the average number of fruits per individual fruits dry weight (DW)/vegetative DW (leaves + stems) per calculated. shoot

Physiological measurements Shoot elongation The following physiological measurements were taken from In C. album, growth of new branches occurs after flowering, each of the 40 marked individuals (10 individuals per 2 generating a bud-scale scar in the stem. In each of the 40 genders per 2 treatments) in February (beginning of flowering marked individuals, ten randomly chosen branches were and coldest month), May (beginning of fruit ripening) and July marked in the last bud scar (10 branches per 10 individuals (end of the growing season and fruit ripening and summer per 2 genders per 2 treatments; n ¼ 400), and the growth of drought) in both study years. newly produced shoots was measured over two consecutive Water potential was measured on one terminal twig per indi- growing periods following Gibson and Menges (1994). vidual using a Scholander pressure chamber (Scholander et al., Measurements were made fortnightly throughout the 2004, 1965). and monthly throughout the 2005 growing periods (February Photochemical efficiency of chlorophyll a (photosystem II) to July), with the aim of studying the effects of inflorescence was measured using a portable modulated fluorimeter bud removal during the current and subsequent growing (MINI-PAM, Walz Effeltrich, Germany) following the modu- periods. To calculate inflorescence bud removal effects on lated pulse-amplitude technique (Bilger et al., 1995). vegetative allocation, the following measurements were con- Calculations followed Schreiber et al. (1995). Maximum ducted on each branch. photochemical efficiency of chlorophyll a (Fv/Fm)was Shoot elongation (SE) of newly produced shoots was calcu- measured on leaves previously adapted to darkness for lated separately for each of the ten marked branches per plant 20 min (three twigs per 40 individuals). The average value as the sum of the elongation values of all the ramifications pro- of the three Fv/Fm measurements was calculated for each indi- duced for each new shoot at each measuring date following vidual (n ¼ 40). Zunzunegui et al. (2002). The mean SE per individual was cal- Both water potential and photochemical efficiency of chlor- culated for each sampling date (10 individuals × 2 genders × ophyll a were measured at solar midday on sunny days, the 2 treatments; n ¼ 40). time of maximum environmental stress, with the purpose of 21 Shoot elongation rate (SErate,mmd ) was calculated as the recording the eventual physiological differences related to increase in shoot length between two consecutive measuring reproductive effort. 992 A´ lvarez-Cansino et al. — Gender-specific costs of reproduction in a dioecious shrub

Data analysis ln (n + 1) for a greater homogeneity in standard deviation and a better fit to the normal distribution. The non-normal Differences in shoot elongation rate (SE ) between males rate variables number of ramifications and number of fruits were and females and between controls and bud-removed plants analysed for differences between treatments and genders by were analysed by means of a repeated-measures ANOVA means of a Mann–Whitney’s U non-parametric test. with gender and treatment as fixed factors. Due to the effect All statistical analyses were performed using SPSS 17 of plant size in determining the patterns of reproductive allo- (SPSS, Chicago, IL, USA) except nested ANOVA, for which cation (Ban˜uelos and Obeso, 2004; Friedman and Barrett, the Statistica 6 (StatSoft, Tulsa, OK, USA) software package 2009) individual canopy size was included as a covariate. In was used. addition, two-way ANOVAs were performed separately for each measurement month, with the aim of detecting the effect of inflorescence bud removal on the different phases RESULTS Downloaded from https://academic.oup.com/aob/article-abstract/106/6/989/140671 by guest on 05 May 2020 of the reproductive season. Although the branches of some species can be considered as A total of 25 154 inflorescence buds was removed from 20 modules or independent units (Tuomi et al., 1982), in order to marked plants; 18 865 (75 %) from male plants and 6289 (25 consider the effect of individual plants and take into account %) from females. Male individuals exhibited an average of 22 possible inter-individual differences in size and/or microsite 3960 + 492 inflorescences m canopy, significantly greater effects, a nested ANOVA was carried out to compare differ- than the values for female plants, with 1148 + 185 inflores- 22 . . ences among treatments of mean shoot elongation (MSE), cences m canopy (F1 ¼ 31 23, P , 0 001). These data are accumulated shoot elongation (SEa) and biomass production consistent with previous results from the authors in studies data. Treatment was considered as the fixed factor and individ- of C. album (Zunzunegui et al., 2006). uals as the random factor nested in treatment. To compare Fruit production in control plants showed an inter-annual 22 differences between genders in each treatment, a nested variation, with lower values in 2004 (1651 + 375 fruits m 22 ANOVA was carried out, with gender as the fixed factor and canopy) compared with 2005 (3855 + 592 fruits m individual as random factor nested in gender. canopy) (F ¼ 8.73, P , 0.05). The possible effect of differences in shoot size on the effect of the experimental treatment on MSE, SEa and biomass production variables was previously dismissed by means of a Effect of inflorescence bud removal on shoot elongation Levene’s test for equality of variance. Homoscedasticity of Bud removal had a significant effect on SErate over the two variance was supported in shoot size for both study periods: growing seasons studied. Repeated-measures ANOVA indi- 2004 (male, F ¼ 0.631, P ¼ 0.428; female, F ¼ 1.866, P ¼ cated the existence of a significant interaction between 0.174) and 2005 (male, F ¼ 0.232, P ¼ 0.858; female, F ¼ month, gender and treatment; thus, the intensity of monthly . . 1 352, P ¼ 0 246). SErate depended both on the treatment and the gender Physiological measurements of water potential and photo- (Table 1). chemical efficiency of chlorophyll a were analysed by means Bud-removed male individuals (Mbr) showed greater shoot of a repeated-measures ANOVA to contrast the interaction of elongation rate during the first flowering period (March to sampling date on the effects of gender and treatment. April 2004; Fig. 1); during the rest of the cycle, the elongation Normality of data was previously checked using a rate was greater in control males (Mc), although differences Kolmogorov–Smirnov non-parametric test. Shoot elongation were not significant. Bud-removed female individuals (Fbr) rate, MSE and biomass production data were transformed by showed greater elongation rates than control females (Fc)

TABLE 1. Results of repeated-measures ANOVA for the effect of treatment and gender on shoot elongation rate (SErate), leaf water potential and maximum photochemical efﬁciency (Fv / Fm)

2004 2005

Variable Comparison MS FPMS FP

Shoot elongation rate Month 0.520 5.963† 0.001*** 0.543 11.460‡ 0.001*** Month × gender 0.539 6.43† 0.001*** 0.728 26.796‡ 0.001*** Month × treatment 0.311 2.486† 0.043** 0.245 3.243‡ 0.021** Month × gender × treatment 0.603 8.345† 0.001*** 0.222 2.846‡ 0.036** Leaf water potential Month 1.25 23.51 0.001*** 0.92 137.27 0.001*** Month × gender 0.006 0.11 0.890 0.482 11.18 0.001*** Month × treatment 0.06 1.127 0.332 0.177 2.58 0.096* Month × gender × treatment 0.035 0.652 0.525 0.122 1.67 0.209 Maximum photochemical efﬁciency Month 0.022 20.46 0.001*** 0.91 126.80.001*** Month × gender 0.000 0.97 0.908 0.05 0.659 0.526 Month × treatment 0.001 6.36 0.572 0.03 0.434 0.653 Month × gender × treatment 0.001 5.46 0.349 0.07 0.966 0.394

F, 2 degrees of freedom, except where indicated otherwise: † d.f. ¼ 7 and ‡ d.f. ¼ 5. Asterisks indicate signiﬁcant differences (* P , 0.1; ** P , 0.05; *** P , 0.01). A´ lvarez-Cansino et al. — Gender-speciﬁc costs of reproduction in a dioecious shrub 993

Flowering Fruiting Flowering Fruiting

Female br F* Female control br** 2·5 c* br** Male br Male control br**

) 2·0 –1 M* c* F* br** br** 1·5 Downloaded from https://academic.oup.com/aob/article-abstract/106/6/989/140671 by guest on 05 May 2020 M* c* 1·0 br** F* F* Bud removal c*** M** c* br*** c** Shoot elongation rate (mm d 0·5 br** c*

0·0 Jan Feb Mar Apr May Jun Jul Feb Mar Apr May Jun Jul 2004 2005

21 F IG. 1. Shoot elongation rate (SErate,mmd ) in male and female individuals [control and bud-removed (br)] throughout the 2004 and 2005 growing periods (mean + s.e.). Upper-case letters indicate significant differences between treatments (M for male plants and F for female plants); lower-case letters indicate significant differences between sexes (c for control treatment and br for bud-removed). Asterisks indicate significant differences (* P , 0.05; ** P , 0.01, *** P , 0.001). The length of flowering and fruiting periods is shown. during fruit production (May to June), although differences with bud-removed males at the end of the growing season, were not significant. In July, once the fruits were formed, Fc though only in the first year. No differences were found showed higher values than Fbr (Fig. 1). In 2005, Fbr individuals between treatments in female individuals (Fig. 2 and showed greater SErate than controls from March to May, while Table 2). Between genders, control males tended to show a in males the differences between treatments were not signifi- higher number of ramifications than females in 2004, although cant (Fig. 1). differences were not significant. Significant differences were Between genders, two-way ANOVA revealed significant found in both treatments in 2005, with male individuals differences in both bud-removed and control individuals in showing a greater number of ramifications than females. both growing seasons. Fbr individuals showed growth rates Accumulated shoot elongation at the end of the growing equal to or higher than Mbr, which were more evident through- period (SEa) was greater in bud-removed female individuals out the second growing period after bud removal (Fig. 1). In in both study years, although one-way nested ANOVA only control individuals, Fc showed higher shoot elongation rates showed significant differences between treatments in 2005 than Mc at the beginning of the two growing seasons. At the (Fig. 3 and Table 3). In male individuals, no differences end of the growing season (July), after fruit formation between treatments were found in any of the study years. period, Fc individuals again showed higher elongation rates Bud-removed female individuals tended to have a higher than Mc in both study years (Fig. 1). SEa than males in both study years, but differences were not One-way nested ANOVA showed significant differences significant (Fig. 3). In control individuals, males tended to between treatments in MSE in both genders (Fig. 2). Control have a greater SEa in both years, with significant differences male individuals had lower MSE than bud-removed males also present in 2005. during most of the 2004 growing season while, in 2005, SEa was lower in both genders and treatments in 2005 com- control males had a greater MSE in March and May. In pared with 2004, due to the severe drought which occurred female individuals, bud-removed plants had greater MSE during this second period. than controls only in 2005. Between genders, in the 2004 growing season (March to May), the nested ANOVA showed lower MSE in male control individuals compared with Effect of inflorescence bud removal on biomass production female controls. No significant differences between genders Inflorescence bud removal showed a significant effect on were found in MSE in bud-removed individuals. In 2005, sig- vegetative DW (leaves plus stems), but only in female individ- nificant differences between genders were found in both treat- uals (Fig. 3). No significant differences were found between ments, with greater MSE values in female individuals. genders on biomass production. Similar to the MSE results, Mann–Whitney’s U non- In females, no significant differences between treatments parametric test showed a significantly greater number of rami- were found at the end of the growing season in 2005, in fications per terminal shoot for control males in comparison fruit DW (Fig. 3), in the proportion of biomass DW devoted 994 A´ lvarez-Cansino et al. — Gender-specific costs of reproduction in a dioecious shrub

Flowering Fruiting Flowering Fruiting

Female br Ramification number Ramification 25 Female control F : 1·74 ± 0·12 number br ± Fbr: 1·34 0·05 Male br F : 1·68 ± 0·2 c ± Male control Fc: 1·16 0·06 M : 1·42 ± 0·07 br ± c** Mbr: 1·51 0·1 20 M : 2·00 ± 1·14 M*** c ± Mc: 1·30 0·1 c** br*** c* br*** M** 15 F* F* c** br*** M* Downloaded from https://academic.oup.com/aob/article-abstract/106/6/989/140671 by guest on 05 May 2020 M*** c*** F* 10 c*** br*** Bud removal M*** c*** c** F** M* Mean shoot elongation (mm) 5 M**

0 Jan Feb Mar Apr May Jun Jul Feb Mar Apr May Jun Jul 2004 2005

F IG. 2. Mean shoot elongation per shoot (MSE) + standard error in both treatments for male and female individuals throughout the 2004 and 2005 growing periods. Mean ramifications number + s.d. is also shown for each treatment and gender. Upper-case letters indicate significant differences between treatments (M for males plants and F for female plants); lower-case letters indicate significant differences between sexes (c for control treatment and br for bud-removed). Asterisks indicate significant differences (* P , 0.05; **P,0.01, *** P , 0.001). The length of flowering and fruiting periods is shown.

TABLE 2. Results of the non-parametric Mann–Whitney’s U test than females in February in the control treatment (P , 0.05; for the effects of treatment and gender on number of Fig. 4A); conversely, in May, female individuals showed ramifications in 2004 and 2005 lower values than males for both treatments (P , 0.01 and P , 0.05, respectively; Fig. 4A). 2004 2005 Repeated-measures ANOVA on the variable maximum Source of variation Comparison UPUPphotochemical efficiency (Fv/Fm) indicated significant differences between months only, with no interactions between Treatment Male 2.00 0.005** 29.50 0.205 genders or treatments (Fig. 4B and Table 1). In 2004, both . . . . Female 19 50 0 498 39 00 0 258 genders showed Fv/Fm values around 0.7 at all sampling Gender Bud-removed 33.00 0.196 17.00 0.022* dates, very close to the optimum, set between 0.7 and 0.8by . . . . Control 12 00 0 916 21 50 0 018* several authors (Bjo¨rkman and Demmig, 1987; Valladares et al., 2000). In February and July of 2005, F /F values Asterisks indicate significant differences (* P , 0.05;** P , 0.01). v m under the physiological optimum were recorded. to fruits (Fbr,1.35 + 0.18, Fc,1.82 + 0.23), nor in the number of fruits (Mann–Whitney’s U ¼ 4365.00; P . 0.05). DISCUSSION Within single female individuals, no differences in biomass production were found comparing fruit-bearing with fruitless Trade-offs between reproduction and vegetative growth branches in any of the treatments (P . 0.05; data not shown). Suppression of reproductive allocation resulted in an increased shoot elongation rate during the reproductive period in both genders, indicating a direct reduction in the somatic cost of Physiological response to inflorescence bud removal reproduction. According to the results, bud-removed female Repeated-measures ANOVA showed the existence of sig- (Fbr) individuals showed higher growth rates than bud-removed nificant month–gender and month–treatment interactions in males (Mbr) in May, when fruits start to form, and during the 2005 (P , 0.1), indicating that monthly variation in leaf whole period in the following growing season (Fig. 1). The water potential was gender dependent and was also affected results demonstrate that the effect of inflorescence bud by suppression of reproduction (Table 1). Accordingly, removal was higher in the gender bearing the greatest repro- one-way ANOVA showed that inflorescence bud removal had ductive effort. a significant effect on water potential in 2005; in February, Throughout the reproductive event, gender differences control male individuals showed lower values than found in the growth rate response to inflorescence bud bud-removed males (P , 0.05; Fig. 4A). Between the removal can be explained through the differences in flowering genders, male individuals showed significantly lower values and fruiting periods in each sexual form. Gender-related A´ lvarez-Cansino et al. — Gender-specific costs of reproduction in a dioecious shrub 995

60 250 SEacumulated 2004 SEacumulated 2005 50 FVB* Leaf biomass Stem biomass 200 Fruit biomass 40 FVB* Biomass (mg) F* 150 30 F* c* c* 100

20 Downloaded from https://academic.oup.com/aob/article-abstract/106/6/989/140671 by guest on 05 May 2020

50 10 Accumulated shoot elongation (mm)

0 0 Female br Female control Male br Male control

F IG. 3. Accumulated shoot elongation (SEa), biomass production (stem, leaves and fruits dry weight) per branch at the end of the growing season. Mean + s.e. Upper-case letter F indicates significant differences between treatments for female plants; FVB indicate differences in vegetative biomass in female plants; lower- case letter c indicates significant differences between genders for control treatment. Asterisks indicate significant differences (* P , 0.05).

TABLE 3. Results of one-way nested ANOVA for the effect of treatment and gender on SEa

2004 2005

Source of variation Comparison MS FPMS FP

Treatment Male 1.134 2.878 0.114 0.013 0.026 0.873 Female 0.241 0.200 0.662 5.043 7.752 0.012* Gender Bud-removed 1.810 1.679 0.211 0.015 0.019 0.892 Control 0.267 0.916 0.507 1.878 6.268 0.023*

F, 1 degree of freedom. Asterisks indicate significant differences (* P , 0.05). variations in growth investment that reflect reproduction represented as reductions in elongation and accumulated timing are common in dioecious species (Obeso, 2002, and shoot elongation (SEa) rates, indicates that both processes are references therein), in which the vegetative growth of female at least partially dependent on the same resource pool individuals surpasses that of males during flowering, while (Nicotra, 1999). shoot elongation of males surpasses that found in the In the present study, inflorescence bud removal is shown to females during fruiting (Delph, 1990). have consequences on the current growing season but also The allocation of resources towards reproduction in control delayed effects on the subsequent growing period, at the end individuals involved a somatic cost that became apparent of which the bud-removed females showed greater SEa than through a reduction in shoot elongation rate, following the the controls (Fig. 3 and Table 3). Delayed effects of reproduc- same trend in the two study periods (Fig. 1 and Table 1). tion on growth have been studied in several species such as This finding is consistent with results obtained by the Siparuna longiflora, in which manipulation of flower and authors in previous years (Zunzunegui et al., 2006). Female fruit production produced an increase in growth of female indi- individuals started to grow earlier than males and their viduals that was noticeable in the subsequent growing period growth rate increased gradually until the middle of May, (Nicotra, 1999). In Aesculus californica, Newell (1991) when it decreased, coinciding with fruit production (Fig. 1). described an indirect expenditure in fruit-bearing branches By the end of the growing season, both sexual forms had accu- that had a delayed effect on reproduction and vegetative mulated equal shoot elongations (Fig. 3). However, during the growth. In other species, such as Lindera benzoin or Salix flowering period of control males (Mc) in 2004 and 2005 and alaxensis, manipulation under natural conditions revealed the of Mbr in 2005, their growth rate was lower than that of female existence of balances between current and subsequent repro- plants (Fig. 1). ductive events, although not with subsequent vegetative These results are consistent with the presence of a trade-off growth (Fox and Stevens, 1991; Cipollini and Whigham, between investments in reproductive biomass and vegetative 1994). biomass, as has been described for other dioecious species The fact that only female individuals showed evidence of (Vasiliauskas and Aarseen, 1992; Gibson and Menges, resource allocation towards growth, when the reproductive 1994). The negative effect of reproduction on growth, investment of the previous period had been suppressed, 996 A´ lvarez-Cansino et al. — Gender-specific costs of reproduction in a dioecious shrub

–1

M* c** –2 br* c*

–3 Downloaded from https://academic.oup.com/aob/article-abstract/106/6/989/140671 by guest on 05 May 2020

Leaf water potential (MPa) –4 Female br Female control Male br –5 Male control A 0·9 B ) m F /

v 0·8 F

0·7

0·6

0·5 Maximum photochemical efficiency (

0·4 February May July February May July 2004 2005

F IG. 4. Leaf water potential (A) and maximum photochemical efficiency (B) at the three sampling dates in 2004 and 2005. Mean values + s.e. are shown. Upper-case letter M with asterisks indicates significant differences between treatments for male plants, lower-case letters indicate significant differences between genders (c, control; br, bud-removed plants). Asterisks indicate significant differences (* P , 0.05; ** P , 0.01). could be explained by the existence of gender-specific differ- organs and water relations to assess compensating mechanisms ences in compensatory mechanisms of reproduction, promoted of reproductive effort in further research. by the greater reproductive effort of the females. Previous The fact that females produced more fruits in 2005 may studies have shown evidence of a longer root system in have reinforced the effect of the inflorescence bud removal female C. album individuals; in the same study site, on vegetative allocation. Control females would have less ´ Alvarez-Cansino et al. (2010), using xylem water isotopic accumulated resources than Fbr and, under higher fruit pro- composition analysis, found evidence of deeper soil layer duction, the delayed effect on vegetative costs would be water extraction by females. Roots act as a resource storage enhanced. This is in accordance with the pattern found by system – mainly for carbohydrates – that can be mobilized Milla et al. (2006), who measured higher biomass and nutri- and allocated to reproduction during the reproductive period ents accumulated in the branches of males of Pistacia lentiscus (Ehrleń and van Groenendael, 2001); a more developed root in the current-year shoots, resulting in a relatively higher level system would thus allow a greater resource storage capacity of resources for a stronger flowering effort in males during the in female individuals. Hence, we conclude that inflorescence subsequent reproduction period. bud removal would imply resource allocation to non- Inter-annual differences in fruit production have been reproductive organs in the subsequent cycle, greater in described in many species (Hoffmann and Alliende, 1984), females compared with male individuals. These results point and the results concur with previous data described for out the importance of considering below-ground storage C. album (Font Quer, 1992; Zunzunegui et al., 2006). A´ lvarez-Cansino et al. — Gender-specific costs of reproduction in a dioecious shrub 997

In addition, 2005 was a drier than average year, which would and high radiation cause a greater physiological stress in the also increase gender dimorphism in physiological responses, gender that invests the greatest effort during flowering. which have been described as being fostered under harsher Gender-related differences found in May 2005, when the environmental conditions (Obeso, 2002). females showed lower leaf water potential values than the Inflorescence bud removal only had an effect on ramification males, were more marked between control plants than in male individuals (Fig. 2 and Table 2), which suggests that between the bud-removed individuals. However, these differ- gender dimorphism in some morphological traits, in addition ences were not correlated with differences in the photosyn- to being genetically conditioned, could also be affected by thetic response, measured as maximum photochemical reproductive status. The effect of inflorescence bud removal efficiency of chlorophyll a (Fv/Fm) (Fig. 4B and Table 1). on ramification and on mean shoot elongation (MSE) could be We could have expected an increase in photosynthesis in explained by the fact that floral meristem production can affect control plants relative to bud-removed individuals, which the development of secondary vegetative meristems through would contribute to the carbon allocated to the production of Downloaded from https://academic.oup.com/aob/article-abstract/106/6/989/140671 by guest on 05 May 2020 hormone production (Lehtila, 2000), that could be present in reproductive structures (Dawson and Ehleringer, 1993; different concentrations depending on the sex. Laporte and Delph, 1996). However, the large reproductive In accordance with the present results in C. album, several investment may not have an effect on the photosynthetic rate authors have explained the greater ramification of male indi- if plants respond by increasing their resource uptake or by viduals of anemophyllous species as an adaptation for using resources more efficiently. Many studies on sexually improved pollen production and exposition (Friedman and dimorphic species show no evidence of a local increase in Barrett, 2009). Male individuals of Simmodsia chinenesis photosynthesis due to reproduction, in which the reproductive (Kohorn, 1994) and of Leucadendron species (Bond and expenditure could be compensated by the reallocation of Midgley, 1988) exhibited shorter internodes as a consequence resources from adjacent branches (Obeso et al., 1998). The of a greater number of inflorescences relative to the females. complementary costs of increasing resource uptake could cor- We consistently found that male C. album individuals, which respond to the costs of production and maintenance of roots or have greater numbers of flowers than the females, had more leaves (Case and Barrett, 2004). ramified branches with shorter internodes (Fig. 2). In conclusion, the initial hypothesis that the effect of inflor- Differences found between treatments in the biomass pro- escence bud removal would be greater in female individuals of duction of female individuals (Fig. 3) confirm the results of C. album, which bear the greater reproductive effort due to shoot elongation and support the existence of a balance in fruit production, is confirmed. The results show evidence for resource investment between reproduction and maintenance differences between sexual forms in somatic reproductive of vegetative growth which, in the case of female individuals, costs and in the presence of delayed somatic costs, indicating would have an effect lasting from the current growing period gender-related differences in resource allocation and storage. to the subsequent period (A˚ gren, 1988; Nicotra, 1999). In addition, it has been shown for the first time that gender- Inflorescence bud removal implied an increase in vegetative related morphological and physiological differences develop biomass production, compared with control individuals in in C. album as a response to different reproductive resource the subsequent growing period (2005), which only took requirements, and are thus dependent on reproductive status. place in females (Fig. 3). However, reproduction in the sub- Long-term manipulative studies are confirmed as a useful sequent period was not affected by the suppression of flower method to assess reproductive costs and to explore the pro- buds, as the same number and dry weight of fruits were cesses underlying the evolution of gender dimorphism in found for both treatments in this study (Fig. 3). plants. This study gives new evidence of the importance of considering roots as essential storage and water control organs that may play a key role in reproductive allocation pat- Physiological responses to inflorescence bud removal terns and in the magnitude of gender dimorphism in sexually Reductions in delayed reproductive costs can be related both dimorphic species. to biomass allocation patterns and to compensation coupled to an increase in photosynthetic assimilation (Case and Barrett, ACKNOWLEDGEMENTS 2004); therefore, we would expect a response of the photosynthetic system to the inflorescence bud removal performed in We thank the Junta de Andalucıá for granting access permits to C. album. Moreover, individuals of the gender and treatment the Donãna Natural Park and to J. Jauregui for helpful comments that invest most in reproduction would be expected to show on the manuscript. This work was supported by the Ministry of greater physiological stress (Verdu´ et al., 2004). Education and Science (F.P.U. AP20021733 to L.A.-C.). According to the present results, inflorescence bud removal affected the water status of male individuals during flowering LITERATURE CITED (February) in the year following that of flower removal A˚ gren J. 1988. Sexual differences in biomass and nutrient allocation in the (2005), with control individuals showing more negative dioecious Rubus chamaemorus. Ecology 69: 962–973. water potential values than bud-removed (Fig. 4A and A´ lvarez-Cansino L. 2009. Gender dimorphism in the dioecious shrub Corema Table 1). Equally, differences between genders during this album at population and biogeographical scales. PhD Thesis, University of Seville, Spain. period were only found in control individuals. These differ- ´ ences could be related to the high photoinhibition found in Alvarez-Cansino L, Zunzunegui M, Dıáz Barradas MC, Paz Esquivias M. ´ 2010. Physiological performance and xylem water isotopic composition males in a previous study (Alvarez-Cansino, 2010), which underlie gender-specific responses in the dioecious shrub Corema confirms that environmental conditions of low temperatures album. Physiologia Plantarum 140: 32–45. 998 A´ lvarez-Cansino et al. — Gender-specific costs of reproduction in a dioecious shrub

Andersson S. 2005. Floral costs in Nigella sativa (Ranunculaceae): compen- Horibata S, Hasegawa SF, Kudo G. 2007. Cost of reproduction in a spring satory responses to perianth removal. American Journal of Botany 92: ephemeral species, Adonis ramosa (Ranunculaceae): carbon budget for 279–283. seed production. Annals of Botany 100: 565–571. Ashman TL. 1994. A dynamic perspective on the physiological cost of repro- Karlsson P, Meńdez M. 2005. The resource economy of plant reproduction. duction in plants. American Naturalist 144: 300–316. In: Reekie EG, Bazzaz FA. eds. Reproductive allocation in plants. Ban˜uelos MJ, Obeso JR. 2004. Resource allocation in the dioecious shrub London: Elsevier, 1–49. Rhamnus alpinus: the hidden costs of reproduction. Evolutionary Kohorn LU. 1994. Shoot morphology and reproduction in jojoba: advantages Ecology Research 6: 1–17. of sexual dimorphism. Ecology 75: 2384–2394. Bilger W, Schreiber U, Bock M. 1995. Determination of the quantum effi- Laporte MM, Delph LF. 1996. Sex specific physiology and source-sink ciency of photosystem II and of non-photochemical quenching of chlor- relations in the dioecious plant Silene latifolia. Oecologia 106: 63–72. ophyll fluorescence in the field. Oecologia 102: 425–432. Lehtila K. 2000. Modelling compensatory regrowth with bud dormancy and Bierzychudek P, Eckhart V. 1988. Spatial segregation of the sexes of dioe- gradual activation of buds. Evolutionary Ecology 14: 315–330. cious plants. American Naturalist 132: 34–43. Leigh A, Cosgrove MJ, Nicotra AB. 2006. Reproductive allocation in a

Bjo¨rkman O, Demmig B. 1987. Photon yield of O2, evolution and chloro- gender dimorphic shrub: anomalous female investment in Gynatrix Downloaded from https://academic.oup.com/aob/article-abstract/106/6/989/140671 by guest on 05 May 2020 phyll fluorescence characteristics at 77K among vascular plants of pulchella? Journal of Ecology 94: 1261–1271. diverse origins. Planta 170: 489–504. Levins R. 1968. Evolution in changing environments. Princeton, NJ: Princeton Bond WJ, Midgley J. 1988. Allometry and sexual differences in leaf size. University Press. American Naturalist 131: 901–910. Matsuyama S, Sakimoto M. 2008. Allocation to reproduction and relative Case AL, Barrett SCH. 2004. Environmental stress and the evolution of reproductive costs in two species of dioecious Anacardiaceae with con- dioecy: Wurmbea dioica (Colchicaceae) in Western Australia. trasting phenology. Annals of Botany 101: 1391–1400. Evolutionary Ecology 18: 145–164. Milla R, Castro-Dıéz P, Maestro-Martıńez M, Montserrat-Martı´ G. 2006. Cipollini ML, Whigham DF. 1994. Sexual dimorphism and cost of reproduc- Costs of reproduction as related to the timing of phenological phases in tion in the dioecious shrub Lindera benzoin (Lauraceae). American the dioecious shrub Pistacea lentiscus L. Plant Biology 8: 103–111. Journal of Botany 81: 65–75. Narbona E, Dirzo R. 2010. Experimental defoliation affects male but not Dawson TE, Ehleringer JR. 1993. Gender-specific physiology, carbon female reproductive performance of the tropical monoecious plant isotope discrimination, and habitat distribution in boxelder, Acer Croton suberosus (Euphorbiaceae). Annals of Botany 106: 359–369. negundo. Ecology 74: 798–815. Newell EA. 1991. Direct and delayed costs of reproduction in Aesculus Dawson TE, Geber MA. 1999. Sexual dimorphism in physiology and mor- californica. Journal of Ecology 79: 365–378. phology. In: Geber MA, Dawson TE, Delph LF. eds. Gender and Nicotra AB. 1999. Sexually dimorphic growth in the dioecious tropical shrub, sexual dimorphism in flowering plants. Heidelberg: Springer-Verlag, Siparuna grandiflora. Functional Ecology 13: 322–331. 175–215. Nicotra AB, Chazdon RL, Montgomery RA. 2003. Sexes show contrasting Delph LF. 1990. Sex-differential resource allocation patterns in the subdioe- patterns of leaf and crown carbon gain in a dioecious rainforest shrub. cious shrub Hebe subalpina. Ecology 71: 1342–1351. American Journal of Botany 90: 347–355. Delph LF. 1999. Sexual dimorphism in life history. In: Geber MA, Dawson Obeso JR. 2002. The costs of reproduction in plants. New Phytologist 155: TE, Delph LF. eds. Gender and sexual dimorphism in flowering plants. 321–348. Heidelberg: Springer-Verlag, 149–173. Obeso JR, A´ lvarez-Santullano M, Retuerto R. 1998. Sex ratios, size distri- Dıáz Barradas MC, Munõz Reinoso J. 1992. The ecology of vegetation of butions, and sexual dimorphism in the dioecious tree Ilex aquifolium the Asperillo dune system, southwest Spain. In: Carter RWG, Curtis (Aquifoliaceae). American Journal of Botany 85: 1602–1608. TGF, Sheehy-Skeffington MJ. eds. Coastal dunes: geomorphology, Popp JW, Reinartz JA. 1988. Sexual dimorphism in biomass allocation and ecology and management. Rotterdam: Balkema, 211–218. clonal growth of Xanthoxylum americanum. American Journal of Botany Dumka D, Bednarz CW, van Iersel MW. 2003. Effect of flower bud removal 75: 1732–1741. on carbon dioxide exchange rates of cotton. Communications in Soil Primack RB, Hall P. 1990. Costs of reproduction in the pink lady’s slipper Science and Plant Analysis 34: 1611–1621. orchid: a four-year experimental study. American Naturalist 136: Ehrleń J, van Groenendael J. 2001. Storage and the delayed costs of repro- 638–656. duction in the understorey perennial Lathyrus vernus. Journal of Ecology Scholander PF, Hammer HT, Bradstreet ED, Hemmingsen EA. 1965. Sap 89: 237–246. pressure in vascular plants. Science 148: 339–346. Font Quer P. 1992. Plantas medicinales. El Diosco´rides renovado. Barcelona: Schreiber U, Hormann H, Neubauer C, Klughammner C. 1995. Assesment of Labor. photosystem II photochemical quantum yield by chlorophyll fluorescence Fox JF. 1995. Shoot demographic responses to manipulation of reproductive quenching analysis. Australian Journal of Plant Physiology 22: 209–220. effort by bud removal in a willow. Oikos 72: 283–287. Tuomi J, Niemela P, Mannila R. 1982. Resource-allocation on dwarf shoots Fox JF, Stevens GC. 1991. Costs of reproduction in a willow: experimental of birch (Betula pendula): reproduction and leaf growth. New Phytologist responses vs. natural variation. Ecology 72: 1013–1023. 91: 483–487. Friedman J, Barrett SCH. 2009. Wind of change: new insights on the Tuomi J, Hakala T, Haukioja E. 1983. Alternative concepts of reproductive ecology and evolution of pollination and mating in wind-pollinated effort, costs of reproduction, and selection in life-history evolution. plants. Annals of Botany 103: 1515–1527. American Zoologist 23: 25–34. Garcıá MB, Ehrleń J. 2002. Reproductive effort and herbivory timing in a Valde´s B, Talavera S, Fernańdez Galiano E. 1987. Flora vascular de perennial herb: fitness components at the individual and population Andalucıá Occidental. Barcelona: Ketres. levels. American Journal of Botany 89: 1295–1302. Valladares F, Wright SJ, Lasso E, Kitajima K, Pearcy RW. 2000. Plastic Gehring JL, Delph LF. 2006. Effects of reduced source-sink ratio on the cost phenotypic response to light of 16 congeneric shurbs from a of reproduction in females of Silene latifolia. International Journal of Panamanian rainforest. Ecology 81: 1925–1936. Plant Sciences 167: 843–851. Vasiliauskas SA, Aarseen LW. 1992. Sex ratio and neighbour effects in Gibson DJ, Menges ES. 1994. Population structure and spatial pattern in the monospecific stands of Juniperus virginiana. Ecology 73: 622–632. dioecious shrub Ceratiola ericoides. Journal of Vegetation Science 5: Verdu´ M, Villar-Salvador P, Garcıá-Fayos P. 2004. Gender effects on the 337–346. post-facilitation performance of two dioecious Juniperus species. Guitiań P, Medrano M, Rodrı´guez M. 1997. Reproductive biology of Functional Ecology 18: 87–93. Corema album (L.) D. Don (Empetraceae) in the northwest Iberian Zunzunegui M, Dıáz Barradas MC, Aguilar Silva FJ, Ain Lhout F, Clavijo Peninsula. Acta Botanica Gallica 144: 119–128. A, Garcıá Novo F. 2002. Growth response of Halimium halimifolium at Hartemink N, Jongejans E, de Kroon H. 2004. Flexible life history four sites with different soil water availability regimes in two contrasted responses to flower and rosette bud removal in three perennial herbs. hydrological cycles. Plant and Soil 247: 271–281. Oikos 105: 159–167. Zunzunegui M, Dıáz Barradas MC, Clavijo A, A´ lvarez-Cansino L, Ain Hoffmann AJ, Alliende MC. 1984. Interactions in the patterns of vegetative Lhout F, Garcıá Novo F. 2006. Ecophysiology, growth timing and repro- growth and reproduction in woody dioecious plants. Oecologia 61: ductive effort of three sexual forms of Corema album. Plant Ecology 183: 109–114. 35–46.